We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.DC

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Distributed, Parallel, and Cluster Computing

Title: Asynchronous Distributed Optimization with Redundancy in Cost Functions

Abstract: This paper considers the problem of asynchronous distributed multi-agent optimization on server-based system architecture. In this problem, each agent has a local cost, and the goal for the agents is to collectively find a minimum of their aggregate cost. A standard algorithm to solve this problem is the iterative distributed gradient-descent (DGD) method being implemented collaboratively by the server and the agents. In the synchronous setting, the algorithm proceeds from one iteration to the next only after all the agents complete their expected communication with the server. However, such synchrony can be expensive and even infeasible in real-world applications. We show that waiting for all the agents is unnecessary in many applications of distributed optimization, including distributed machine learning, due to redundancy in the cost functions (or {\em data}). Specifically, we consider a generic notion of redundancy named $(r,\epsilon)$-redundancy implying solvability of the original multi-agent optimization problem with $\epsilon$ accuracy, despite the removal of up to $r$ (out of total $n$) agents from the system. We present an asynchronous DGD algorithm where in each iteration the server only waits for (any) $n-r$ agents, instead of all the $n$ agents. Assuming $(r,\epsilon)$-redundancy, we show that our asynchronous algorithm converges to an approximate solution with error that is linear in $\epsilon$ and $r$. Moreover, we also present a generalization of our algorithm to tolerate some Byzantine faulty agents in the system. Finally, we demonstrate the improved communication efficiency of our algorithm through experiments on MNIST and Fashion-MNIST using the benchmark neural network LeNet.
Comments: 37 pages, 4 figures. Related to our prior work on Byzantine fault-tolerance distributed optimization in redundancy in cost functions (doi:10.1145/3382734.3405748 and arXiv:2101.09337)
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC)
Cite as: arXiv:2106.03998 [cs.DC]
  (or arXiv:2106.03998v1 [cs.DC] for this version)

Submission history

From: Shuo Liu [view email]
[v1] Mon, 7 Jun 2021 22:40:18 GMT (1531kb,D)

Link back to: arXiv, form interface, contact.